Evaluation of the Heatwave Plan for England Final report

(1)

PIRU Publication 2019-24

Lorraine Williams, Bob Erens, Stefanie Ettelt,

Shakoor Hajat, Tommaso Manacorda and

Nicholas Mays

(2)

(3)

November 2019

Lorraine Williams, Bob Erens, Stefanie Ettelt,

Shakoor Hajat, Tommaso Manacorda and

Nicholas Mays

(4)

Ahmed at the Royal College of Nursing (RCN) for their help with the nurse survey.

Finally we would like to thank the charitable organisations who helped set up the focus groups, in particular Sharon Frankland and colleagues from Contact the Elderly.

This research is funded by the National Institute for Health and Research (NIHR) Policy Research Programme, conducted through the Policy Research Unit in Policy Innovation and Evaluation (reference 102/0001). The views expressed are those of the author(s) and not necessarily those of the NIHR or the Department of Health and Social Care.

(5)

Acronyms

British Social Attitudes (Survey) Committee on Climate Change Clinical Commissioning Group Chief Executive Officer Community Risk Registers Cold Weather Plan for England Department of Health and Social Care Environmental Audit Committee

Emergency Planning Resilience and Response Focus Group

Frontline Practitioner Government Office Region General Practitioner Housing Association Heatwave Plan for England Health Research Authority Health and Well-Being Board Health and Well-Being Strategy

Information and Communications Technology Information Technology

Joint Strategic Needs Assessment Local Authority

Local Health Resilience Partnership Limiting Long-Standing Illness Local Resilience Forum

London School of Hygiene and Tropical Medicine National Centre for Social Research

National Health Service

National Severe Weather Warning Service Office of National Statistics

Public Health England

Policy Innovation and Evaluation Research Unit Particulate Matter

Royal College of Nursing Risk Registers

Statistical Package for the Social Sciences Sustainable Transformation Plan

World Meteorological Organisation BSA

CCC CCG CEO CRR CWP DHSC EAC EPRR FG FLP GOR GP HA HWP HRA HWBB HWBS ICT IT JSNA LA LHRP LLSI LRF LSHTM NatCen NHS NSWWS ONS PHE PIRU PM RCN RR SPSS STP WMO

(6)

Key findings and recommendations

Evaluation of the Heatwave Plan for England

• The Heatwave Plan for England (HWP) aims to protect health and reduce harm from severe hot weather. It is a good practice guide underpinned by a heat-health watch alert system. The alert system is managed by Public Health England (PHE), who commission the Met Office to provide the forecast for the alerts. In the event of an upcoming period of severe weather where regionally defined temperature thresholds are predicted to be breached, heat-health alerts are cascaded to the Cabinet Office and all health, social care and public services, including NHS providers and local authorities, to enable them to implement their local protection plans.

• This evaluation examined the contribution of the HWP to protecting the health of the population during hot weather by conducting: 1) a time-series analysis to establish the relationship between hot weather and adverse health outcomes; 2) case studies of local implementation of the HWP in five areas in England, along with a national survey of nurses in hospital, community and care home settings;

and 3) a survey of the general public to explore whether people protect themselves and others by following the advice set out in the HWP.

Epidemiological relationship between hot weather and health

• The relationship between temperature and mortality, and between temperature and emergency hospital admissions (as indicators of the health impact of hot weather), suggests that hot weather in England is associated with an increase in deaths and emergency hospital admissions.

• There is no evidence that general summertime relationships between temperature and mortality and between temperature and emergency hospital admissions have changed substantially in the years since the introduction of the first HWP in 2004.

• Since the largest number of excess deaths and hospital admissions associated with heat take place outside of heatwave alert periods, this raises questions about the appropriateness of current threshold levels as well as the need to place more emphasis on general preparedness strategies as represented by levels 0 and 1 of the HWP.

Implementation of the Heatwave Plan for England

• Some interviewees noted that there could be sub-regional variation in maximum temperatures within Met Office regions, with some areas (e.g. on the coast) being less likely to experience severe hot weather even when other areas within the same region exceed the alert threshold. This led to some local authorities ignoring regional alerts, and to possibly underestimate current and future risks, as they rarely experienced temperatures that exceeded alert thresholds.

• Local heatwave plans were closely aligned with the national HWP, particularly in the heat-health alert system guidance offered in the Plan. However, heatwaves were often assessed as lower risk than other weather-related hazards (such as floods and cold weather) and were often given a lower priority in planning.

(10)

• Heatwave planning was largely seen as an exercise in emergency preparedness and focused on ‘warning and informing’ through the alert system, rather than as a strategic objective of long-term public health and environmental planning.

• The role of Clinical Commissioning Groups (CCGs) in planning and implementing local heatwave plans was not clear; in some areas CCGs were reported to be taking a key role in planning and co-ordinating the health response, while in others they were said to be acting in a more supportive role, with NHS England taking the lead.

• Emergency planners, mainly in local authorities and acute trusts, said that they adopted a ‘wait and see’ approach, employing professional judgment before escalating actions during a heatwave. Some noted that plans may not work as well if a heatwave alert comes over a weekend, when relevant managers may not be at work.

• Many frontline staff, including nurses surveyed, reported to be unaware of any local heatwave plans, and unfamiliar with the HWP guidance, but most said they knew what to do to protect their patients and clients ahead of and during heatwaves.

However, not all the actions stated by frontline staff during the 2017 and 2018 heatwaves were appropriate or in line with HWP guidance, and many reported taking few or none of the recommended HWP actions during an alert.

• Many frontline nurses surveyed said that they struggled to protect their patients during heatwaves, reporting their organisations to be often ill prepared for severe heat events. Many said that they were working in difficult and challenging conditions: often in old and poorly designed buildings not well adapted for climate change; a lack of funding and resources to implement many HWP actions; and often poor working conditions with inflexible organisational policies.

• During alert periods, it was reported to be difficult to reach all high-risk groups, especially those who might be ‘below the radar’ of health or social services, such as people with social care needs who do not qualify for means-tested social support, agricultural labourers and homeless people. Managers said that they tended to rely on health information for the general public to reach these groups.

• General practices may often be well placed to support these high-risk patients, as they tend to have better access to vulnerable groups missing to other NHS providers and local authorities, but their role in the HWP was not clear as they do not have contractual duties for emergency planning, and there are concerns about whether they would have capacity.

• Unless there was a heat-related ‘major-incident’, few mechanisms were said to be in place to monitor activities during and following a heatwave alert, so managers were not able to formally assess how well their organisations performed during the alert period.

(11)

Risk awareness and behaviour of the general public

• Most adults in England do not consider themselves at risk from hot weather, including a majority of those considered to be ‘vulnerable’ (adults with certain health conditions or older adults aged 75 and over). Rather, the general public, especially those aged under 45, tends to hold positive views about hot weather.

• While most of the public recognise the effectiveness of some actions to protect themselves from hot weather (e.g. drinking cool fluids, staying out of the sun between 11am and 3pm), they do not appreciate the effectiveness of other actions (e.g. keeping exposed windows and curtains closed and avoiding alcohol).

• Knowledge of the effectiveness of behaviours was related to the actions people actually took during the June 2017 heatwave. However, this appears to be mitigated by peoples’ perceptions of risk, so that individuals who do not perceive themselves to be at risk will not even take protective actions they know to be effective.

• Younger adults were more likely than older people to report experiencing hot weather related health symptoms during the 2017 summer period. This may be a reflection of their greater under-estimation of the risks of hot weather and reduced likelihood of taking protective behaviours as a result. It may also be that the effects of hot weather on older people (e.g. cardiovascular or lung conditions) are less easily identified as resulting from the heat.

• Despite national and local promotion of heat protection messages during the summer months for many years now, knowledge of some protective behaviours among the general public is still poor (e.g. about closing exposed windows during the day), and the publicity/advice appears not to be reaching some vulnerable groups (e.g. adults who report being in bad health).

Recommendations

• PHE to consider to organise, and possibly rename, the ‘Heatwave Plan’ to a

‘Summer Health Protection Plan’ to acknowledge that preparation is needed to reduce the risk of adverse health effects during temperatures below the current heat-health alert thresholds.

• PHE should ensure that more encouragement is given to local managers and frontline staff to improve their awareness of the HWP and to take stronger action in hot weather, recognising that heat-health harm begins to occur as soon as temperatures rise to average summer levels and well below the levels that trigger alerts.

• PHE to review procedures in the Plan to identify and provide preventive services to vulnerable people who are not routinely in contact with health or social care providers, such as older people who live on their own and transient groups such as agricultural labourers.

• PHE and the Local Government Association to review the capacity and capability of local authorities and other health and social care partner organisations, including those providing voluntary and community services, to implement protective actions

(12)

arising from the HWP. This should include considering how primary and community care organisations could work together better to provide guidance, support and protection to vulnerable individuals and groups before and during extreme heat events.

• PHE to review the advice provided to local authorities and health and social care providers on planning for hot weather. This would include providing evidence- based recommendations to hospitals, care homes and similar facilities on air conditioning; improving staff welfare during severe weather events, particularly for those working in areas that are difficult to keep cool; and prioritising HWP awareness through mandatory training for all healthcare staff.

• PHE to revise public health advice/publicity in order to: improve public awareness of the risks of hot weather to health; to enable a realistic self-assessment of risk among different population groups; tailor messages to the information needs and media usage of different population groups, including younger and older adults;

and increase knowledge of the effectiveness of those protective behaviours of which many people are unaware (e.g. closing windows and curtains in direct sunlight during the day).

(13)

The evidence that the world is warming is now unequivocal. Many countries are experiencing higher summer temperatures and a greater frequency of heatwaves, including the UK. There is consistent evidence that exposure to high temperatures, such as those during a severe heat event, directly increases mortality and morbidity, particularly amongst the most vulnerable groups such as infants and young

children, pregnant women, older people and those with chronic conditions such as cardiovascular and respiratory disease (Public Health England, 2015b, Hajat et al., 2010, Kim et al., 2019).

Following the severe heatwave in 2003, which accounted for over 2,000 excess deaths in the UK, and climate change predictions of more frequent and increasingly hotter summers in England, the Government introduced its first Heatwave Plan for England (HWP) in 2004. The aim of the HWP is to protect the population from heat-related harm to health by planning to ‘prepare for, alert people to, and prevent, the major avoidable effects on health during periods of severe heat in England’ (Public Health England, 2015a p.7). Much like other similar public health schemes adopted in many European countries at the same time, the HWP incorporates a heat-health warning system which triggers short-term protective measures when severe hot weather is forecast. However, whilst similar heat-health warning and adaptive advice systems have been shown to reduce heat-related mortality, research in this area is limited, particularly for those targeting vulnerable groups (Toloo et al., 2013, Lowe et al., 2011).

Since 2004 the HWP has been occasionally updated to take account of new research, and was most recently refreshed in 2018 (Public Health England, 2018b).

In 2016, the Department of Health and Social Care (DHSC) commissioned the Policy Innovation & Evaluation Research Unit (PIRU) at the London School of Hygiene & Tropical Medicine (LSHTM) to conduct an independent evaluation of the implementation and potential effects of the HWP. The evaluation was conducted between January 2017 and October 2018.

The evaluation addressed the following three questions:

1. Has the introduction of the HWP in 2004 had any effect in terms of reducing morbidity and mortality?

2. To what extent, if any, has the HWP informed local decisions on management of heat-related health risk and response?

3. Is the general population aware of the risks of heat and overheating buildings, do they change their behaviour in hot weather as a result of hearing heat alerts/advice in line with the HWP, and do they take any actions to prevent potential effects of hot weather (e.g. adapt their homes)?

The evaluation comprised a mixed method study involving four components:

1. A time series analysis of health data linked to hot weather at the level of

Government Office Regions as well as the 44 areas covered by the Sustainability and Transformation Plans (STPs), to characterise heat-health relationships and trends over time.

2. A national survey of knowledge, attitudes and behaviour of the general population during heatwaves.

3. Case studies of heatwave planning and implementation in selected local areas in England over a 12-month period.

4. A national survey of nursing staff in hospital, community and care home settings on their awareness of the HWP and actions taken during heat-health alerts.

1. Introduction

and background

(14)

This report is aimed at a variety of stakeholders including policy makers as well as local planners and implementers of the HWP in the English health and social care system. It presents a broad range of findings over the four strands of work, drawing a number of conclusions and recommendations from these findings.

1.1 The Heatwave Plan for England (HWP)

The HWP presents itself as a good practice guide setting out ‘what should happen before and during periods of severe heat in England. It sets out what preparations both individuals and organisations can make to reduce health risks and includes specific measures to protect at risk-groups’ (Public Health England, 2015a: p.7). The plan sets out actions which should be taken to reduce harm, including a national heat-health alert service as well as heat-health guidance and advice to the general public, local communities and public service providers including local authorities, NHS, social care, voluntary groups and other public agencies. The stated aims of the HWP are ‘to raise public awareness of the dangers of excessive heat to health and to ensure that health, social care and other voluntary and community organisations and wider civic society is prepared and able to deal with a heatwave when it comes so as to protect the most vulnerable’ (Public Health England, 2015a: p7).

There is no universally agreed definition of a heatwave. In an effort to develop a global standard, the World Meteorological Organisation’s Task Team on Definitions of Extreme Weather and Climate Events suggested the following definition:

‘A marked [sic] unusual hot weather (maximum, minimum and daily average) over a region persisting at least two consecutive days during the hot period of the year based on local climatological conditions, with thermal conditions recorded above given thresholds.’ (TT-Dewce, 2016: p10)

On their website the UK Met Office defines a heatwave as ’an extended period of hot weather relative to the expected conditions of the area at that time of year’ (Met Office, 2018). Temperature thresholds for heatwaves, however, are not the same as thresholds defined to trigger a heat-health alert. Heat-health alerts, limited to England, have been created by the Met Office in conjunction with PHE and targeted primarily at health professionals and emergency planners. This uses threshold maximum daytime and minimum night-time temperatures that vary by region, when the average temperature thresholds set at 30°C by day and 15°C overnight occur for at least two consecutive days (Public Health England, 2015a). These temperature levels were chosen based on the balance between the risk to health and the risk of sending out alerts. They set the point at which services might be expected to ‘ramp up’ their activities in accordance with the severe nature of the weather.

1.2 The heat-health alert system

The heat-health alert service is a core part of the HWP. It operates in England between 1st June and 15th September. The HWP sets out the rationale for

introducing a system of hot weather alerts as ‘unlike cold weather the rise in mortality as a result of very warm weather follows very sharply, within one or two days of the temperature rising’ and there is only a short window for effective action once the heatwave starts, so that ‘advanced planning and preparation is essential’ (Public Health England, 2015a: p6). On receiving an alert, responsible authorities are

(15)

expected to cascade the alert within their organisations, and to commissioners and providers of health, public health and social care services (Appendix 1). Five alert levels are described in the Plan (Figure 1.1). Specific preventative and protective actions, triggered by regional temperature thresholds issued by the Met Office, are recommended when temperature thresholds are forecast (level 2) and reached (level 3) in local areas (Public Health England, 2015a). Alert levels are colour coded to standard emergency warning systems, from blue (level 0) to red (level 4).

Public Health England provides further support during the alerts through monitoring outputs such as information on excess mortality and morbidity. This would include any increase in heat-related illness such as an increase in calls to NHS 111 or GP consultations.

In addition to the heat-health alerts, the HWP provides heat-health information in the form of leaflets, posters, checklists and action cards aimed at the general public;

health and social care professionals responsible for supporting vulnerable people;

teachers and professionals responsible for children; and those working in care homes.

The HWP also links to a separate publication for health and social care professionals that sets out the evidence in support of the heat-health guidance (Public Health England, 2015b).

Heat-health alerts issued in 2017 and 2018

In 2017 a level 2 heat-health alert was issued on June 16th and a level 3 on June (18- 20th). The mean average temperature in England was 16 degrees centigrade (Public Health England, 2017). In 2018 level 2 heat-health alerts were issued in June (26- 28th), July (9th) and (12-19th) and August (1st). Level 3 heat-health alerts were issued in July (2-8th) and (23-27th) and August (6th) and a record mean average temperature for England was recorded of 17.2 degrees centigrade (Public Health England, 2018c).

1.3 Brief overview of relevant studies

Effects of heat on health

The negative effect of hot weather on the health of the population is well established in England and elsewhere. Previous studies suggest that adverse health effects occur, particularly in those most vulnerable to hot weather such as people aged 75 and over, infants and young children, people with severe physical or mental illnesses, and those with pre-existing medical conditions (Nayak et al., 2018, Thompson et al., 2018, Bassil and Cole, 2010, Green et al., 2010, Wang et al., 2012, Stafoggia Figure 1.1 Heatwave alert levels

Level 0 Long-term planning – All year

Level 1 Heatwave and Summer preparedness programme – 1 June – 15 September

Level 2 Heatwave is forecast – Alert and readiness – 60% risk of heatwave in the next 2 to 3 days Level 3 Heatwave Action – temperature reached in one or more Met Office National Severe Weather

Warning Service regions

Level 4 Major incident – Emergency response – central government will declare a Level 4 alert in the event of severe or prolonged heatwave affecting sectors other than health

Reproduced from the Heatwave Plan for England (2015) p.14.

(16)

et al., 2006, Arbuthnott and Hajat, 2017). Age risk, especially older age, is likely to be due to a diminished ability to thermoregulate, increased likelihood of co- morbidities or conditions that limit or affect adaptive behaviours during hot weather (such as dementia), or use of medications which affect thermoregulation (Public Health England, 2015b, Arbuthnott and Hajat, 2017). Those working outdoors during summer months, such as agricultural workers, are also identified as high-risk (Spector et al., 2016). Other identified high-risk groups include the homeless and rough sleepers, as these groups are more likely to have poor overall physical and mental health and co-morbidities such as respiratory conditions which are made worse during heatwaves and/or as a result of substance dependencies (e.g. drugs and alcohol) which can increase heat stress. These groups are also viewed as a high- risk of heat morbidity and mortality from other factors such as social isolation and exposure to the environment (Public Health England, 2015b). A recent review found that high ambient temperatures had a range of negative effects on mental health, particularly in relation to increased suicide risk (Thompson et al., 2018).

Awareness of risk and preventive behaviour

Following the 2013 heatwave in the UK, an online survey of 1497 people found that high-risk groups, such as people living in dense urban areas that are susceptible to the ‘heat island effect’, and those on low incomes, were less likely to take preventive measures during a heatwave than other groups (Khare et al., 2015). A review of the literature investigating public risk perception and behaviour found that, among those people who were aware of an extreme heat episode, few changed their behaviours;

a lack of self-perception of being ‘vulnerable’, and confusion about the right action to take were the main reasons for not changing behaviour (Bassil and Cole, 2010).

However, this review explored mostly US, Canadian and other European studies, with only one UK study included. In addition to the Bassil and Cole findings, one author’s explanation for the discrepancy between heat-health knowledge and action is that risk is minimal and not life threatening, even among those likely to be more vulnerable to heat (Burchell et al., 2017). It has also been noted that people may confuse the harms from UV radiation with harms from heat, thus underestimating their risk of dehydration or other heat-related risks (Wolf et al., 2010, Burchell et al., 2017). Older people in particular do not perceive themselves at risk from hot weather (Abrahamson et al., 2009, Wolf et al., 2010, Bittner et al., 2014).

Studies also show that preventive behaviours were highly variable (Bassil and Cole, 2010, Wolf et al., 2010, Khare et al., 2015, Toloo et al., 2013, Waldock et al., 2018) and that both physiological and psychological factors have been shown to play a part.

For example, Waldock and colleagues found that older people can have reduced perceptual awareness of their thermal environment, meaning that they may not feel uncomfortable enough in a hot environment, and therefore may be less likely to implement cooling behaviours, such as seeking shade or removing excess layers of clothing (Waldock et al., 2018). Another study found that heat protection measures (such as those triggered during heat-health alerts) were not perceived as warnings, but rather as positive news by the general population since they triggered fond memories of long hot summers in the past, and this led to a reduction in protective measures taken by people as they sought to enjoy the hot weather (Lefevre et al., 2015).

There is also some uncertainty about whether heat-health protection methods are effective. A recent systematic review of prevention methods, including heat-health behavioural interventional advice such as those within the HWP, found methodological challenges and inconclusive evidence of effect (Boeckmann and Rohn, 2014).

(17)

Implementation and effectiveness of the HWP

Heat-health warning and response systems are in operation in a variety of European countries. Though similar to the HWP there are key differences to allow much comparison (Lowe et al., 2011). There is some evidence to support the effectiveness of these systems in reducing heat related mortality (Toloo et al., 2013), but research in this area is limited.

England has had a heatwave plan in place since 2004 and several studies have looked at how it has been implemented and its effects. Planning and preparing for heatwaves sit within a suite of competing severe weather and other local emergency events, and are often given low priority. A recent small survey of members of Local Resilience Forums (those responsible for emergency planning in local areas) showed that heatwaves were the ‘severe weather event’ they felt least well prepared for (nine out of 17 interviewed), but reasons were not provided (Cox and Crouch, 2017). Another concern is that responsibilities and tasks of stakeholders are not well described in the HWP, leading to potential confusion about local roles (Abrahamson and Raine, 2009).

There are only a few studies examining the views and actions of those implementing the HWP, such as health and social care managers and practitioners. When the HWP was first introduced, health and social care professionals found it useful in preparing for heatwaves, but they were unsure whether the Plan resulted in any actions that would help those most vulnerable to hot temperatures (Johnson and Bickler, 2007).

Several recent studies suggested that local implementers tended to give low priority to heatwave planning and were insufficiently familiar with the HWP, particularly frontline health and social care staff (Wistow et al., 2017, Abrahamson and Raine, 2009, Boyson et al., 2014, Gupta R et al., 2016, Woodward, 2014), though one author concluded that frontline hospital staff in some care settings were aware of the dangers of heat and felt able to provide appropriate care during hot weather (Boyson et al., 2014). Another recent study found that some staff in care homes were unaware of the risks heatwaves posed for older residents and that care home managers need to better prepare their homes to deal with the effects of climate change (Gupta R et al., 2016). Another study showed that within hospitals, communication between managers, who had good knowledge of local heatwave planning, and frontline staff, who lacked such knowledge, was an issue during heatwaves (Boyson et al., 2014).

The HWP advises health and social care organisations to be proactive by ensuring that they have systems in place to identify and contact their most vulnerable populations in the event of a heatwave. However, a recent study showed that heat-related health behaviour of frontline staff was reactive rather than anticipatory, the authors concluding that there is a need to share information about vulnerable people more effectively between organisations responsible for implementing the HWP (Wistow et al., 2017).

Other studies highlighted difficulties in identifying and reaching vulnerable population groups during heatwaves and the importance of involving local community groups, in addition to voluntary ‘emergency responders’ (e.g. local branches of the Red Cross), in developing and implementing local heatwave plans (Burchell et al., 2017).

These findings resonated with the evaluation of the Cold Weather Plan in England (introduced in 2011) that found that vulnerable populations were difficult to identify and community resilience was lacking (Heffernan et al., 2018). The authors of this study also noted that leadership of local cold weather planning was highly variable and that GPs were difficult to engage in cold weather preparedness.

(18)

Little is known about the impact of the HWP on long-term preparedness for hot weather and few studies have examined the effects of national government climate change initiatives at local government level (Tompkins et al., 2010). A recent report by the Government’s Environmental Audit Committee recommended that the Government should be more proactive to ensure safe and resilient homes and support local authorities to better ‘green’ their local areas (House of Commons, 2018). Overheating in homes remains a risk for a large number of English households as buildings, particularly newer builds, lack heat resilience (Beizaee et al., 2013, Taylor et al., 2015, Mavrogianni et al., 2014). This is also the case for some care homes whose buildings are less able to cope with climate change and who lack investment strategies to adapt (Gupta R et al., 2016). It is also estimated that 90 percent of UK hospital buildings are prone to overheating (ARCC, 2015). Climate analysts predict that heatwaves are likely to contribute to more deaths in the future in England and, as part of any future plans, there will be an increasing need to adapt existing and new buildings to better withstand heat, whilst reducing carbon emissions through low energy design strategies (Committee on Climate Change, 2014, The Lancet, 2018, Short et al., 2012, ARCC, 2015).

More robust evidence is needed on the role of wider structural and social

determinants of adverse outcomes from exposure to hot weather, e.g. the type and state of private housing (Kovats and Bickler, 2012) as well as temperature regulation in hospitals (Carmichael et al., 2012, Short et al., 2012) and care homes.

(19)

2.1 Epidemiological relationships between hot weather and health

The first objective of the evaluation of the HWP involved an epidemiological assessment of retrospective data to characterise the nature of the relationships between high temperatures and indicators of health at the levels of Government Office Regions (GOR) and Sustainability and Transformation Plan (STP) areas. Changes in heat-health relationships since the introduction of the HWP are also assessed as well as potential links between annual heat-risk and previous wintertime experience.

Data

The health data consisted of:

• All deaths occurring in England during 1st Jan 1981 – 31st Dec 2015, obtained from the Office for National Statistics.

• All emergency hospital admissions occurring in England during 1st April 1997 – 31st March 2012, obtained from NHS Digital. These data were already available from a previous evaluation of the Cold Weather Plan (Hajat et al., 2016) and a new data application was not made in the interests of time.

Each health outcome was aggregated by date to create a time-series of the daily number of events for each GOR and for each of the 44 STPs. Separate series were also created by age and disease groups.

The exposure data consisted of daily mean, minimum and maximum temperatures for the same time periods. These were recorded by Met Office land surface stations.

For each measure, one composite series was created for each region by combining data from stations recording measures on at least 75% of days during the study period and using a previously published imputation method to deal with missing values (Armstrong et al., 2011). On average, 20 stations contributed data to each regional series. A similar approach was used to create composite series for each STP however all stations were included regardless of the amount of missing values in order to maximise available data.

Statistical analysis

Time-series regression analysis was used to characterise the short-term (i.e. day- to-day) associations between temperature and health indicators. For each series of deaths or emergency hospital admissions, slow-changing seasonal patterns in the health counts (unrelated to temperature) and any secular trends were controlled for using splines of time, with seven degrees of freedom per year of data analysed.

Spline functions are a series of polynomial curves (usually cubic) joined together to flexibly model patterns in a time series of health data. Indicator terms were used to model any day-of-week effects. The relationship between temperature and health indicators was then assessed graphically, again using spline functions. As effects of high temperatures are mostly immediate, impacts distributed up to two days following exposure are quantified, although longer lags were also assessed. In general, the graphical relationships indicated a gradual increase in the risk of a heat-related health event once daily temperatures increased above certain threshold levels. For quantification purposes, therefore, a linear threshold model was used, whereby there is assumed to be no risk at temperatures below the threshold value, and a linear relationship between temperature and risk of heat-related health event above the threshold. To objectively identify the heat threshold for each region, statistical model diagnostics were compared between models with threshold values fixed at different temperatures. Effects were estimated for just the summer months defined as the months of June to September.

2. Methods

(20)

The time-series models were used to (i) compare temperature-health relationships before and after the introduction of the HWP, (ii) explore the alert thresholds currently used in the HWP, and (iii) assess intra-summer variation in heat vulnerability. It was not possible to conduct a direct comparison of mortality impacts on alert days compared to non-alert days or days of similar temperatures before the advent of the HHWS as there were too few days of such extreme temperatures, for example in London there were only two days when our daily maximum temperature series reached the alert threshold value of 32°C. Furthermore, we did not have information on the whole period in which days an actual alert was called by the Met Office based on their forecast values.

For the final epidemiologic component, in order to explore possible links between summertime vulnerability and previous winter burdens, we also quantified heat risk for individual summers and correlated this against the mortality experience of the previous winter. Some studies have observed that a low winter mortality burden may result in a higher than expected mortality burden in the following summer due to the pool of people at risk remaining large (Ha et al., 2011). As both annual summertime and wintertime mortality has been decreasing over time, series were de-trended using spline functions prior to assessment of correlations.

Results are illustrated for London and West Midlands although the patterns were similar in all regions.

2.2 Local implementation of the Heatwave Plan for England:

Case studies

Overview of case study design

A longitudinal multiple case study approach, employing qualitative research methods, was used to investigate how the HWP was implemented in five local areas over a period of one year. The case study design used is well-established in research in many public policy settings, including healthcare (Exworthy et al., 2011, Crowe et al., 2011). It allows critical events, policy development and programme-based service reforms to be studied in detail in their ‘real life’ context (Yin, 2009). Our case studies focused on how the HWP was interpreted and implemented locally through local plans and subsequent activities, using interviews with managers and frontline staff, the analysis of relevant policy documents and observations of meetings and workshops, enabling in-depth investigation of local processes and activities before, during and after a heatwave.

A logic model (Weiss, 1998) (Figure 2.1) was developed by the team at the early stage of the research, to enable the research team to visually conceptualise the intervention and its processes. This enabled framing of specific evaluative questions by looking at local context, implementation and expected outcomes. The model was based on the current HWP (Public Health England, 2015a). It illustrates the process by identifying the broad activities, outputs and intended outcomes from the HWP.

(21)

Figure 2.1 Logic model of the Heatwave Plan for England*

* Note: Level 4 (emergency response) is minimally represented in this model for reasons of simplicity and clarity.

Inputs

NHS Trusts and LAs’ Local

Heatwave Plans reviewed,

updated and in place

Health, social care and public service providers

prepare for imminent heatwave

Protective actions taken by health and social care and public service providers

and general public

National emergency

actions NHS and social care/public service

providers and CCGs

NHS England National and Regional teams Local Authorities/PH directors/LRFs/LHRPs

General public Public communications

TV/radio/newspaper Public Health England

Heatwave Plan for England

Department of Health & Social Care and other government departments

Civil Contingencies Secretariat

Met Office cascades heatwave alerts

Vulnerable/high-risk patients and people protected from effects of heatwaves.

General public aware of heat-health and taking protective measures.

Minimal impact of heatwave to health, social care and public funded services

Level 2 Level 3

Level 1 Level 4

Guidance on protective actions and long-term planning for heatwaves Activities

Organisations People

Outputs

Intended outcomes

(22)

Site selection

Five local authority areas were purposively selected as case study sites to allow for variation in geography and demography. Criteria for selection were derived from the research literature, epidemiological data, and discussions with officials at PHE and the DHSC. This was further refined by reference to further regional data . Sites were selected to include a mix of urban, semi-urban and rural areas and three broad English regions: North, Midlands and South. They also represent differences in exposure to heat and weather-related activity patterns, such as the number of visitors engaging in outdoor recreational activities including summer festivals, and the number of agricultural workers in areas in which agriculture constitutes a significant part of the local economy.

Sites also differed in their proportion of older people as a group identified in the HWP as vulnerable to hot weather. One London borough was included as a site, as London has been identified in previous studies as being particularly at risk of heatwaves.

Recruitment of case study sites

The evaluation team contacted the CEO of each local authority selected to ask for their agreement to participate in the evaluation (Appendix 2). Recruitment of sites was supported by a letter of support from PHE (Appendix 3). If they agreed, CEOs were asked to provide the name and contact details of senior staff responsible for heatwave planning in their organisation who would act as the main ‘key’ contact for the evaluation team.

Identification and recruitment of informants in sites

Those leading on heatwave planning in each local authority, ‘key contacts’ for the evaluation e.g. directors of public health or emergency planners with a lead on heatwaves or business continuity, were invited to participate in an interview. An information leaflet, detailing the purpose of the study and their involvement (Appendix 4), and a consent form (Appendix 5) were provided along with the study invitation.

After providing consent, the key contact agreed to direct the evaluation team to relevant staff within the local authority and in other local organisations involved in heatwave planning, which typically formed Local Health Resilience Partnerships.

We recruited staff involved in planning and managing the response to hot weather, as well as staff working at the frontline with at-risk groups in each site. A participant information sheet (Appendix 4) outlining details of the research and their participation was sent prior to obtaining consent.

Managers were recruited directly via the key contact. Managers then helped to identify and recruit frontline staff either by asking them directly, providing their names to the evaluation team, or by posting recruitment leaflets (Appendix 6). Interested staff responded by contacting a member of the research team for further information by email or telephone. A gift token of £25 was provided to frontline staff to compensate them for their time.

Conducting interviews

Semi-structured interviews were carried out using separate interview schedules for managers and frontline staff to reflect their roles in planning and implementation (Appendix 7). Managers were interviewed in person at their place of work or over the phone. Frontline staff were interviewed over the phone. Interviews lasted between 30 and 80 minutes, were audio recorded with consent and transcribed verbatim.

Following the level 3 alert in June 2017, a short follow-up interview was conducted with the key contact (or a deputy) over the phone, to ask about any action taken after

(23)

the alert, efforts to monitor actions and any reflections on the process. In spring 2018, further follow-up interviews were carried out with a sub-sample of staff interviewed in spring 2018 to explore any local learning from the previous heatwave planning cycle and any actions taken in preparation for the 2018 summer.

Documentary data and observations of meetings

The case study analysis also drew on relevant documents, including local strategic and operational documents relating to severe weather or heatwave planning;

Community Risk Registers (CRR); Joint Strategic Needs Assessments (JSNA) and Health and Well-Being Strategies (HWBS); local climate change and environmental strategy documents; and any public health guidance relating to heatwaves provided by the local authorities or its partners. Documents were identified through internet searches on relevant websites of the participating local authorities, local NHS organisations and others. In addition, interviewees were asked to point us to any documents that might be relevant to this study. Documents were analysed using the themes identified during the analysis of the interviews.

Across three of the case study sites we observed one Local Resilience Forum planning meeting, one community resilience workshop and visited one general hospital to note measures taken to improve climate resilience.

Analysis

An iterative approach was employed to data collection and analysis throughout the evaluation. Interviews were analysed thematically using the Framework Method for applied policy research (Ritchie and Spencer, 1994). The original research questions, as well as new ideas generated inductively from the data, influenced the process.

Members of the research team familiarised themselves with the data by reading transcripts and making notes. Three of the research team independently coded the first few transcripts for a priori and emergent themes. Following discussion, an initial analytical framework of emergent codes and descriptors was agreed and applied to a further set of transcripts. The analytical framework was further refined to incorporate new and refined codes, and a final analytical framework was agreed by the team.

The agreed framework was then applied to each transcript using NVivo 11 software and charted within framework matrices for each of the case studies, whereby summarised interview data could be read across codes and cases (participants).

Thematic analysis was undertaken by the research team reviewing the summaries within the matrices and making connections between categories and participants.

Documentary data were analysed descriptively by searching, listing and summarising content in the documents relating to heat or heatwaves using agreed search terms.

NVivo 11 software was used to aid the process. Detailed field notes of meetings and workshops observed were used to provide context to the analysis of interviews.

2.3 Local implementation of the Heatwave Plan for England:

National survey of nurses

The views and experiences of frontline staff were obtained through a web survey of nurses working in hospital, community and care home settings carried out in September and October 2018. This was about one year after most of the interviews for the case studies were conducted, and the nurse survey questionnaire asked about the 2018 summer period, during which there were a number of heat-health warnings throughout the country, and was generally much warmer than the 2017 summer covered by the case studies.

(24)

An email invitation (Appendix 13) was sent on 14 September 2018 to all nurses identified on the Royal College of Nurses (RCN) membership list identified as working in hospital, community or care home settings in England. Student nurses were excluded, as were those working in academia. The questionnaire was completed online, using the Qualtrics survey platform (www.qualtrics.com/uk).

In all, 2697 completed questionnaires were available for analysis (some returns were excluded due to an insufficient number of questions being completed). The final sample included nurses working in:

• Hospital setting (n=1942)

• Community setting (n=518) which includes community health services, primary care, school services, etc.)

• Care home setting (n=237) which includes nursing and residential care homes.

The questionnaire topics covered: background characteristics of nurse participants and their place of work; nurses’ awareness and knowledge of the HWP; actions taken by nurses during heat-health alerts over the 2018 summer period; and how well the nurse participants, and their organisations, are prepared for protecting patients during periods of severe hot weather. The questionnaire is included in Appendix 14.

Analysis

The nurse survey data were ‘cleaned’ by the PIRU research team and the data were analysed using SPSS v23. The analysis presents survey results by setting (hospital, community, care home) and by nurse role (manager, frontline).

Free text responses to the questions on ‘organisations’ preparedness for hot weather’

and ‘any additional points on heatwave planning’ were extracted by setting (hospital, community, care home), saved as Word files and uploaded to qualitative software (NVivo 11) for coding and thematic analysis using the Framework Method (Ritchie and Spencer, 1994).

2.4 Public knowledge and behaviour: Survey and focus groups

Survey

A survey of 1878 members of the general public in England (aged 18 and over), living in private residential addresses, was carried out by National Centre for Social Research (NatCen) in August and September 2017. The sample was drawn from members of NatCen’s random probability panel, which involves largely web-based surveys, but also includes telephone interviews with panel members who do not have internet access in order to provide coverage of the whole population. Of the 3153 panel members in England invited to participate in the survey, the achieved sample included 1633 web interviews and 245 telephone interviews, giving an overall survey response rate of 60%.

NatCen used a sequential mixed mode design, with panel members first invited (by various methods including email, text and post) to complete the survey questionnaire online. Panel members who had not completed the online questionnaire after two weeks were then contacted by telephone (if phone numbers were available). This ensures that panel members who do not have access to the internet, or who may have literacy or language problems with a written questionnaire, are still able to participate. A £5 gift card was sent as a ‘thank you’ to those who participated.

Fieldwork lasted for one month, from 24 August to 24 September.

(25)

Further details of the NatCen probability-based panel can be found in (Jessop C, 2018).

Questionnaire

Draft survey questions were initially specified by the PIRU research team, and these were subsequently modified and refined in discussion with NatCen. The questions covered: attitudes to hot weather; effectiveness of various actions for protecting people from heat; adaptations to home that lend protection from heat; awareness of any hot weather advice or publicity during the summer heatwave in June 2017;

actions taken to reduce harm from heat (for themselves or for other vulnerable people they know); and whether they suffered any ill effects from hot weather during that period. The majority of the questions came from previous surveys that looked at the public’s views on heat protection messages and measures, and on their changes in behaviour (Public Health England, 2016; Lefevre et. Al., 2015). The questionnaire is included in Appendix 12.

The questions on hot weather and the summer heatwave were part of a survey which included modules of questions on other topics for other NatCen clients.

Non-response and weighting

Since not everyone invited to participate in a survey does so, non-response weighting is used to try to minimise any bias introduced by differential response among

population sub-groups (e.g. men versus women, younger versus older age groups, etc.). For surveys using the NatCen panel, non-response can occur at three stages:

firstly, for the survey which is used to recruit the panel (i.e. the British Social Attitudes Survey (BSA), which involves a face-to-face interview with a probability sample selected throughout Great Britain); secondly, refusal to join the panel at the end of the BSA interview; and thirdly, non-response for particular panel surveys.

NatCen calculates a weight to account for non-response at each of the three stages, with the final weight being the product of these three weights. Logistic regression models are used to derive the probabilities of response of each panel member, and the weight is computed to be the inverse of the probabilities of response. The weight adjusts for non-response using a number of variables such as region, household type, education level, internet access and social class. Further details about the weighting and survey methods for this specific survey are available in a Technical Report written by NatCen, which can be obtained from the PIRU research team.

Table 2.1 shows how the estimated profile of the population (column a) compares with the weighted survey sample (column b) (at the time of panel recruitment in 2015/16) for several key socio-demographic variables. It also shows the profile of the achieved panel sample (column c) at the time of data collection for the current survey (summer 2017). While columns (b) and (c) are similar, the latter is updated for people who may have moved from one region to another, changed their type of job or household, etc.

(26)

Table 2.1 Socio-demographics: population estimates compared with weighted survey estimates at time of panel recruitment, and profile of weighted sample at time of interview (summer 2017)

(a) Population

estimate – England (BSA)

%

(b) Sample profile at time of panel recruitment (after weighting)

%

(c) Sample profile at

time of survey (after weighting)

% Gender

Male 48 48 48

Female 52 52 52

Age

18-24 11 12 9

25-34 17 17 17

35-44 17 17 17

45-54 18 18 17

55-64 14 15 16

65+ 22 20 22

Region

North East 5 5 5

North West 13 13 13

Yorkshire & The Humber 10 10 10

East Midlands 9 9 8

West Midlands 10 10 10

East of England 11 11 11

London 16 15 15

South East 16 16 16

South West 10 11 11

Social grade

Managerial & professional 38 39 41

Intermediate 12 14 14

Small employers & own account workers 9 7 8

Lower supervisory & technical 8 8 8

Semi-routine & routine 28 27 29

Household type

Single person household 17 16 17

Lone parent 4 4 4

2 adults (no children) 36 36 35

2 adults (with children) 21 21 23

3+ adults (no children) 15 15 16

3+ adults (with children) 7 7 6

(27)

Analysis

The survey data were ‘cleaned’ by NatCen and provided to the research team as an SPSS dataset. Data analysis was carried out by PIRU researchers using SPSS v23.

The analysis presents survey results by socio-demographic variables including:

gender, age group, ethnic group, Government Office Region (GOR), highest educational qualification, household type, urban or rural residence, longstanding disability and self-reported general health.

Results are also provided for a number of ‘vulnerable’ groups, including (unweighted bases are shown in brackets):

• Individuals aged 75 years and over (75+) (143)

• Individuals aged 75+ who live alone (69)

• Individuals aged 18-74 with a limiting longstanding illness (18-74 with LLSI) (260)

• Individuals aged 18-74 with self-reports of bad/very bad health (18-74 with bad health) (135)

While young children are also recognised as another at-risk group, since the survey did not collect data on the age of children living in participants’ households, it was not possible to identify participants who were looking after young children.

Focus groups

Four focus groups about how people coped in hot weather were carried out in three different geographical regions and towns in England: two in central London; one in the South East; and one in the Midlands. Participants were purposively selected to include those identified potentially as most at risk during heatwaves: largely older people, particularly those over age 75 and living on their own, including those with health conditions likely to be made worse by heat (Public Health England, 2015a: p14).

Participants were recruited through national and local voluntary organisations supporting people from this group, one being a national charity organising monthly tea parties

Table 2.1 Continued (a)

Population estimate – England (BSA)

%

(b) Sample profile at time of panel recruitment (after weighting)

%

(c) Sample profile at

time of survey (after weighting)

% Economic activity

Full time education 5 4 4

Paid work 57 58 54

Unemployed 5 5 5

Retired 23 21 20

Other 11 11 17

Tenure

Owned/being bought 64 64 63

Rented (LA) 10 8 9

Rented (HA/Trust/New Town) 7 8 6

Rented (Other) 18 18 17

Other 1 1 3

(28)

for older people (over age 75) living on their own; others were recruited through local community luncheon clubs for older people. Table 2.2 presents the characteristics of the focus groups.

Recruitment

National charity: The charity was approached by one of the researchers by telephone and email to seek their agreement to help with recruitment. The charity was selected as it organises monthly tea parties in regions across England for small groups of older people aged 75+ who are dealing with loneliness and social isolation. Volunteers for the charity offered to ‘host’ the tea party in their own homes, as well as helping with transporting guests. Once the charity had agreed to help, information about the research including the recruitment leaflet (Appendix 8) was sent to the head office and disseminated regionally. Two of the charity’s area organisers agreed to help with recruitment. The organisers’ communicated our request (via the recruitment leaflet) to their hosts and guests verbally and by email. Those interested were invited to request further information about the evaluation and the focus group in the form of a participant information sheet (Appendix 9). Once a group had agreed to participate (initial consent was required from the host as well as the guests), a date was set to hold the focus group during a tea party and the focus group was convened during the event. In both focus groups, all participants were living independently and some were familiar with others in the group as they had met during previous tea parties.

Luncheon clubs: Three established luncheon clubs for older people were approached by one of the researchers by telephone and email with details about the research, requesting help with recruitment. Luncheon club organisers agreeing to help (two) were provided with further information including a recruitment poster. One focus group was convened in a luncheon club run by a local church which participants attended typically on a weekly basis. The manager of the luncheon club was approached by the researcher to see if any members would like to participate in the focus group. A poster and an information sheet about the research was distributed to the members two weeks prior to the focus group being held. Information about the focus group was publicised in the club centre on their notice board. The researcher attended a week before the planned focus group date to meet with interested members and answer any questions they might have.

Information sheets and consent forms were distributed (Appendices 9 and 10).

All participants were provided with a gift token of £25 to compensate them for their time.

Table 2.2 Characteristics of focus groups Focus

Group Area How recruited Date

held

Attendees

Male Female Age range Total

1 South

East National charity organising

monthly tea party Dec 2017 1 4 5 > 75 5

2 Midlands National charity organising

monthly tea party April 2018 1 5 6 > 75 6

3 London Local luncheon club

(church) May 2018 2 4 6 > 75 6

4 London Local luncheon club

(community) May 2018 1 7 7 > 75

1 65-74 8

Total 5 20 25

(29)

Process

On the day of the focus group, the facilitator (researcher) provided a copy of the information sheet (enlarged for accessibility) and explained the process face-to-face, restating assurances of anonymity and confidentiality. The group had an opportunity to ask any questions before providing their full written consent. The interview was recorded only with consent from all group members, which was given in all cases.

The aims and conduct of the focus group were agreed and people were asked to introduce themselves for the record. The facilitator used a topic guide (Appendix 11) to frame the discussion with probes and prompts as necessary. The topics included general opinion and attitudes to heat and hot weather, the participant’s own and other’s heat-health behaviour, including any identified risks and coping strategies.

Participants were provided with a selection of current health promotion leaflets/

posters such as the HWP’s ‘Beat the Heat’ campaign and Age UK’s guidance on Staying Cool in a Heatwave (Public Health England, 2018a, Age UK, 2017).

Analysis

Interviews were audio recorded and transcribed verbatim by the researcher/facilitator.

A reflective diary was kept, and detailed observational notes were written-up

immediately following each focus group. A thematic framework approach to analysing the data was used (Richie & Spencer 1994), whereby themes were developed from the research questions as well as from the transcript. Familiarisation with the data was undertaken by the researcher personally transcribing the recorded data in tandem with field notes and observational data. Transcripts were read and re-read a number of times and then coded thematically, mapped on a chart and interpreted. IT software to enable the process (NVivo11) was used.

2.5 Patient and public involvement

Three lay Research Advisors were recruited via PIRU’s collaboration with the Quality and Outcomes of Person-centred Care Research Unit (QORU), another DHSC funded Policy Research Unit (PRU) which has a Public Involvement Implementation Group that supports and provides public involvement in research projects across the two PRUs. This Group has recruited a pool of 20-30 Research Advisors who can be called upon to provide input into particular projects.

A briefing note was circulated to the Group. Key tasks were to advise on the design and quality of fieldwork documents and to comment on reports and papers as a result of the evaluation and other dissemination activities. Documents were sent to the Advisors electronically for comment/feedback at various points in the study, such as in the development stage of interview and focus group topic guides and questionnaires, and in draft sections of reports. Their feedback enabled the research team to further refine and improve the documentation for participants (Appendices 4, 5, 6, 8, 9 and 10), interview questions, focus group topic guides (Appendices 7, 11 and 12) and the draft chapter on public attitudes, awareness and behaviour related to hot weather (Chapter 5).

2.6 Ethical approval

Health Research Authority approval to conduct the research was given on 27th March 2017. The London School of Hygiene and Tropical Medicine’s Research Ethics Committee gave the study a favourable opinion on 21st March 2007 (Ref 12004-1).

(30)

3.1 Comparison of temperature-health relationships before and after introduction of the HWP

Temperatures have risen during the study period. For example, in London the average daily summertime maximum temperature was 18.0°C between 1981-1989, 18.6°C between 1990-1999, 19.4°C between 2000-2009 and 20.0°C between 2010-2015;

in the West Midlands over the same time periods it was 16.0°C, 16.4°C, 17.3°C and 18.1°C respectively.

Figures 3.1 and 3.2 show the seasonally-adjusted regional relationships between year-round daily mean temperature and daily mortality and between year-round daily mean temperature and daily emergency hospital admissions respectively. In each graph, the centre line represents the estimated temperature-health relationship, and the lines either side are 95% confidence limits. For mortality, the relationships are shown separately for the 12 years before and 12 years after the HWP was introduced in 2004. For the shorter emergency hospital admissions series, relationships were considered separately for the 6 years before 2003 and 9 years after. The extreme summer of 2003 was excluded from comparisons as relationships could have been heavily influenced by this unusual year. With both health outcomes, the graphs show very little change in the relationship in any of the regions since the plan has been in operation. In interaction models, there was no significant difference in the heat effect (steepness of the slope) between the two time-periods.

3. Epidemiological relationships between hot weather and health

Lead author:

Shakoor Hajat

Evaluation of the Heatwave Plan for England Final report

PIRU Publication 2019-24

Lorraine Williams, Bob Erens, Stefanie Ettelt,

Shakoor Hajat, Tommaso Manacorda and

Nicholas Mays

November 2019

Lorraine Williams, Bob Erens, Stefanie Ettelt,

Shakoor Hajat, Tommaso Manacorda and

Nicholas Mays

Acronyms

Contents

Key findings and recommendations

Evaluation of the Heatwave Plan for England

Epidemiological relationship between hot weather and health

Implementation of the Heatwave Plan for England

Risk awareness and behaviour of the general public

Recommendations

1. Introduction

and background

1.1 The Heatwave Plan for England (HWP)

1.2 The heat-health alert system

1.3 Brief overview of relevant studies

2.1 Epidemiological relationships between hot weather and health

2. Methods

2.2 Local implementation of the Heatwave Plan for England:

Case studies

2.3 Local implementation of the Heatwave Plan for England:

National survey of nurses

2.4 Public knowledge and behaviour: Survey and focus groups

2.5 Patient and public involvement

2.6 Ethical approval

3.1 Comparison of temperature-health relationships before and after introduction of the HWP

3.

Epidemiological relationships between hot weather and health